Deproblematization as an Enrichment of Framing Theory: Enhancing the Effectiveness of an Awareness-Raising Campaign on Child Poverty

ABSTRACT

Entities wanting to communicate purposefully should have insight into the different frames and counter-frames that give meaning to an issue. Therefore, this research introduces a Framing Counter-framing Theory (FCT). A conceptual distinction is made between frames that define an issue as a problem and frames that deproblematize it. An experimental study (N = 1,000) was conducted in Belgium regarding the effects of an awareness-raising campaign on child poverty. It demonstrated that using deproblematizing frames can render such a campaign more effective. There was an increased willingness to donate in the condition in which the campaign used deproblematizing frames as counter-framing strategy. Furthermore, the results highlight how the internal coherence of a deproblematizing frame can be disrupted by priming an alternative problematizing frame mitigating the intended effect. The theoretical and practical implications of these findings and FCT are discussed.

Introduction

Issues are not intrinsically controversial, but only become so as a result of a political and social process (Fowler et al., 2012, p. 172). A great many actors in the social arena – called frame sponsors (Carragee & Roefs, 2004) – have the daily task of formulating problems in a professional manner. Of course, governments, interest groups and nonprofits may define problems differently, starting from ideological frames and by presenting specific narratives, to propose their own solutions for issues (Crow & Lawlor, 2016). Therefore, the question whether and how a subject is defined as a public policy issue is of strategic significance to the goals of many types of organizations (see Heath & Palenchar, 2009). It goes without saying that strategic communication is central to this, as it can determine the success and failure of the formulated objectives (Zerfass et al., 2018).

For certain organizations and communicators, defining an issue as a problem may really be necessary, in order to draw attention to it. For others, deproblematization may be the preferred option, for example, because the dominant representation is stigmatizing for specific groups in society. However, deproblematization does not necessarily mean that something is not problematic in a normative sense. The deproblematization of euthanasia, for example, can be judged as problematic, for instance, because it is considered a light-minded solution. Others, however, would welcome a definition of euthanasia as an expression of the absolute autonomy of a human being (Van Gorp et al., 2021).

Importantly, the solution to an issue might change depending on the way the issue is framed. A frame formulates an answer to the question people ask themselves when they are confronted with a situation they do not yet fully comprehend: “What is it that’s going on here?” (Goffman, 1974, p. 8). A frame is, as it were, brought into a story ‘from the outside’. Once a frame is available in a given culture whereby people are familiar with it, it becomes engrained in the memory and may be used to give meaning to specific issues (Chong & Druckman, 2007). Since the quality of a solution is related to the quality of the problem formulation, it may be that a problem can only be solved after it has been reframed, or as Pretz et al. (2003, pp. 4–5) argued: “Before a path to a solution is found, ill-defined problems often require a radical change in representation.” This compels all entities that want to communicate purposefully to pay special attention to the different ways in which frames contribute to issue definitions and solutions. Accordingly, frames can help organizations to realize their goals (Haydu, 1999).

Some problems force themselves upon us, others must be discovered, or even created, and other issues can become so normalized that they are no longer perceived as problematic (Jamrozik & Nocella, 1998). It is only when something is experienced as a problem that attention can be paid to it, policymakers may tackle it and financial resources can be allocated to resolve it. Immigration (Quinsaat, 2014), child poverty (Platt, 2005), and climate change (Bolsen & Shapiro, 2018) are examples of issues that are defined as problematic in certain countries, by particular actors, within a certain time period and to a greater or lesser extent. Also, with regard to poverty, one can state that westerners living in poverty would be defined as poor compared to the social norms in a consumer culture, but would not necessarily be considered poor in a low-income country (Hamilton et al., 2014).

The focus in this article is on the interplay between frames that problematize and frames that deproblematize child poverty. Deproblematization in relation to poverty can take different forms. Poverty can be seen as a sign of detachment from the material world, or from a consumer society. Alternatively, growing up in poverty may be perceived as a school for learning how to deal with setbacks, how to stand up for oneself, and how to develop one’s talents. Another possibility is that poverty is defined as a problem, but not the person in poverty, which points to the importance of destigmatization (Hamilton et al., 2014). For example, in soup kitchens, homeless persons try to deal with the poverty stigma associated with it, in part by distancing themselves from the other who is “just a loser” (Nichols, 2020, p. 589). Social service providers must therefore be aware of what exactly they define as the central problem on which they focus their activities. As such, actors that concentrate on poverty (see e.g., Bebbington et al., 2010) might redefine the issue as being about injustice and inequality.

Poverty is regularly analyzed from personal and societal levels of causes and solutions (e.g., Kim et al., 2010). If the people affected are made less personally responsible for the poor conditions they live in, the stigma will logically decrease as well. One way to question the personal responsibility of poverty is to look at it from the prism of the child, as a symbol of innocence, purity, and honesty. Targeting children living in poverty can increase the impact of policy, for example, because a child who receives support in obtaining a degree will enjoy the benefits of this later. Therefore, also Belgium, the country this study focuses on, has followed the international trend of concentrating primarily on child poverty since the turn of the millennium, because of a firm belief in the possibilities of breaking the cycle of generational poverty (Vandenbroeck & Van Lancker, 2014).

Seen from a communication point of view, using the image of the vulnerable child is a strategic choice, as it has a strong appeal in recruiting support and charity (Swift, 1995). The child helps to circumvent the perception that poverty is the result of one’s own actions and that one is responsible. This results, however, in what we call the ‘problematization paradox’: Putting the focal point on child poverty may be effective, but also suggests that parents living in poverty are not competent to care for their children. One criticism, therefore, is that the focus on child poverty may obscure the fact that all poor people should experience benefits from social progress (Mestrum, 2011). These are considerations that should play a role in the strategic communications of government or public interest groups that engage with poverty.

Communicating strategically about an issue, we believe, involves a full understanding of the frames that problematize it and frames that deproblematize it. This study’s goal is to formulate a framing–counter-framing theory (CFT). Especially deproblematizing frames could mitigate the disadvantages of problematization, such as stigmatization and a fatalistic view of a possible solution to an issue. Deproblematizing frames then act as counter-frames to the dominant, problematizing framing.

Counter-framing

Counter-framing is a concept that originated from social movement theory (SMT). It refers to each “attempt to rebut, undermine, or neutralize a person’s or a group’s myths, version of reality, or interpretative framework” (Benford, 1987, p. 75). Waller and Conaway (2011) also used the term for indicating actions aimed at reversing the dominant framing. Counter-framing strategies can incite collective action, providing they are able to challenge that dominant view (Noakes, 2005). In these conceptualizations, frames necessarily precede counter-frames in time: the counter-frame is a response to an earlier frame (Chong & Druckman, 2013, p. 2).

Defining counter-framing as challenging the dominant framing poses several difficulties for both the researcher and an organization seeking to communicate strategically. First, there is the question of how it is possible to determine which framing is dominant? This requires an intensive deductive analysis of the prevailing discourse. Second, the dominant position can change and is context dependent. In comparison with other European countries, Belgium can be characterized as a country where perceptions about the characteristics of poverty indicate a high level of personal involvement with the topic, with addiction and laziness being considered prime explanations for poverty, and fewer people placing responsibility with society (Da Costa & Dias, 2014). Third, from a strategic and normative standpoint, it may not be the dominant perspective that needs to be countered, but a minority perspective. For example, one can intuitively tell that the idea that poor people are responsible for their own social position is problematizing and stigmatizing, even when this is not the most prevailing idea in public discourse. Fourth, there are no overarching characteristics of “counter-frames” in SMT, except for the fact that they are a response to an earlier frame. This means that if a frame sponsor formulates an answer to a successful counter-frame, they have created a counter-frame to a counter-frame. Both are subsumed under the category of “counter-frame” although they offer opposing perspectives.

In this context, it is relevant to consider what dynamic pattern may emerge, as the deproblematizing framing or the problematizing framing of an issue may dominate at any given time. It is indeed possible for deproblematization to precede problematization, as it was the case with, for instance, nuclear technology (Gamson & Modigliani, 1989; Nisbet & Huge, 2006). This raises the strategic and normative question of which counter-frame should then be used, a problematizing frame or deproblematizing frame? This again emphasizes the importance of considering the problematizing nature of a frame.

How do frames that deproblematize an issue become effective?

Unexpected approaches

Framing represents a potentially powerful communication effect, as it combines its own internal logic with the cultural resonance of the frame, by which the members within the same culture recognize and perceive the frame as natural and familiar (Gamson, 1992; Snow & Benford, 1992). How, then, could frames that deproblematize an issue become more effective than the problematizing frames? On the one hand, deproblematization does not fit well in a (media) logic which states that issues only deserve attention when they are seen as problematic. On the other hand, deproblematization might come across as an unexpected approach to an issue, in contrast to the problematizing frames about child poverty that are already plentiful in public discourse (e.g., Hamilton et al., 2014; Kim et al., 2010). Therefore, deproblematizing frames can be subject to thoughtful consideration.

Furthermore, a deproblematizing frame that functions as a counter-frame is more likely to resemble something as a gain-frame (with a focus on gains) than a loss-frame (with a focus on losses), because a deproblematizing frame will make the ‘risks’ associated with an issue are less salient. Its embedded solution is more within reach when the problem is less pronounced. Based on the assumption that most people’s perception of their own poverty risk is low, a ‘gain’-frame, and by extension a deproblematizing frame, could be more successful in countering a dominant perspective (Quick & Bates, 2010). However, if deproblematization indeed resembles a gain-frame, it is essential that the message can intensify the desired emotional response (Nabi et al., 2020). In the case of poverty, this would be a gain frame that expresses hope, for example. The conceptualization of frames and counter-frames used here, however, goes beyond the distinction between gain and loss frames. As can be noted in a frame-matrix (Gamson & Lasch, 1983), emphasis frames are much more argumentative (see Supplementary Online Appendix C).

The type of solution

Compared to SMT (Benford & Snow, 2000), our conceptualization argues that all frames include a diagnostic (i.e., a problem definition) and prognostic component (i.e., a solution), albeit the proposed solution may be of a very different nature. If poverty is a ‘stumbling block’, sensitizing the public that, if society does not reduce the thresholds that disadvantaged children experience, their problems will only worsen, causing damage to society as a whole (e.g., due to drug abuse). The emotion appealed to would then be fear. For a counter-frame that deproblematizes, the solution is of a different nature. For instance, if it is stated that a child in poverty has a core capacity to grow up to be someone who is able to take care of themselves (further referred to as The seed) then a positive efficacy appeal speaks from this. Or if an appeal is made to solidarity as the solving capacity of society, then the collective efficacy is addressed (e.g., Donohoo et al., 2018). If someone perceives the collective possibility of a society to tackle a problem as sufficiently large, this can also result in a strengthening of individual self-efficacy (Bandura, 1997). For example, research shows that violence in certain parts of a metropolitan city can be reduced if the social cohesion within a neighborhood is strong enough and it goes hand in hand with the will to fix it (Sampson et al., 1997). All of this culminates in a first research question:

RQ1: To what extent can deproblematizing frames contribute to the effectiveness of a message related to child poverty?

A clash of arguments

Sniderman and Theriault (2004) were the first to study counter-framing intensively. They did this based on the justified conclusion that frames, as the organization of ideas in an issue, seldom occur in isolation in a message. It is more likely that people are confronted with a “clash of arguments” (p. 146), so with competing or complementary ways of defining an issue. They demonstrated that citizens tend to endorse the deproblematizing frame more if it is presented to them separately than if it is offered in combination with a frame that matches their own convictions. Indeed, quantitative measures show that competitive framing cancels out the effects of individual frames (e.g., Aklin & Urpelainen, 2013; Brewer & Gross, 2005; Hansen, 2007; Wise & Brewer, 2010).

It is important to underline, however, that the notion of dual framing does not necessarily mean that the frames contradict each other. Multiple frames can also complement or reinforce each other. As such, the juxtaposition of multiple frames can produce different effects than what would be expected from seeing the individual effects of the frames (Shah et al., 2004), because the public is more motivated to evaluate the position of the frames when multiple perspectives are present in a message (Chong & Druckman, 2007; Rowling et al., 2015). The question remains if all combinations of frames are equal. The research that has investigated the latter question yielded different results. Wise and Brewer (2010) concluded that some combinations of frames are not able to offset the effects of the individual frames. On the other hand, Aklin and Urpelainen (2013) concluded that it does not matter which frame is chosen, because the effects are cancelled out when the frame is faced with competition. However, they also suggest that future research should further investigate the vulnerability of different frames to counter-framing. Because problematizing frames tend to dominate public discourse, mentioning them explicitly – even if it is to contradict them – may strengthen that problematizing view (see Lakoff, 2014). This implies that, regarding child poverty, a deproblematizing frame is expected to be more effective when it is applied in isolation than when it is combined with a problematizing frame. Regarding child poverty, ‘effectiveness’ can then include perceptions of responsibility, which should fall less on the person in poverty; more support for welfare, which has been found to be associated with more social attributions of responsibility (Iyengar, 1990); and a higher willingness to donate to charity. The focus on the child may increase the willingness to donate (Micklewright & Schnepf, 2009), as the attitude toward helping others has also increased (D’antonio, 2014). However, once the association that a child is innocent is eliminated, that willingness might also drop. This leads us to the following hypothesis:

H1: Deproblematizing frames applied in isolation will (H1a) decrease the amount of responsibility placed on the person living in poverty, (H1b) lead to more positive attitudes towards welfare, and (H1c) increase willingness to donate to charity to a greater extent compared to when deproblematizing frames are combined with a problematizing frame.

A frame especially contributes to the quality of internal consistency in a message. The psychological mechanism that subsequently comes into play on the receiver’s side is that particular elements in a message activate the corresponding schema. Based on cognitive schema theory (Wood et al., 2018), it can be stated that there is also an internal logic on the receiver’s side which prevents someone from weighing up schemata against each other. When it comes to FCT, it is crucial to see if and how a counter-frame can break through the internal cohesion of a schema, and thus motivate the audience to reflect on the issue. As such, putting the child at the center of the debate on poverty is a form of counter-framing. It can break through the social representation of poor people as responsible for their own social position. This leads to a next research question:

RQ2: Can a deproblematizing frame break through the internal logic of a problematizing frame, and can it persuade the public not to think of poverty in a problematizing way?

Personal experience as a moderator

The strength and direction of framing effects, and ultimately of framing as a theory of media effects, depends on a set of moderating variables (for an overview, see Lecheler & De Vreese, 2019). An individual-level moderator that we consider to be decisive for framing in a media-centered model is the audience’s direct and personal experience of an issue (Baden & Lecheler, 2012; Nabi et al., 2020). Earlier research (Iyengar, 1990; Kim et al., 2010) showed that media framing affects the assignment of responsibility and solutions for poverty, but it did not take specific account of issue involvement. The literature shows that high involvement increases the chance that a message will be processed more systematically. Yet, depending on whether more pro-attitudinal or more counterarguments are generated, acceptance of the message increases or decreases respectively (Petty & Cacioppo, 1979). Donovan and Jalleh (2000) found only a positive framing effect for participants with low involvement but no effect for those with high involvement. Also, other related research (e.g., Luong et al., 2019; Vyncke & Van Gorp, 2018) stressed the role of issue involvement regarding framing effects, which demonstrate that people with a high level of involvement are the least susceptible to framing. The associated hypothesis is:

H2: Deproblematizing frames lead to stronger changes in (H2a) responsibility attribution, (H2b) attitudes towards welfare, and (H2c) willingness to donate among on those with low involvement, compared to those with high involvement.

Methodology

Sample

Respondents were drawn from an on-line opt-in panel of a polling agency. They were motivated via a financial incentive. We set quotas for language (Dutch and French-speaking), age, gender, and social class. All respondents gave informed consent. An instructional manipulation check was added in the survey, asking participants to pick a specific response option to demonstrate that they had read the instructions (Oppenheimer et al., 2009). Participants who failed to follow the instruction were filtered out. Using a weighting coefficient, the sample was made representative of the Belgian population regarding age, gender, and social class. All statistics reported in this article make use of this weighted sample.

The mean age of the sample was 48.10 (SD = 15.76), with the youngest participants being 18 years old and the oldest aged 79. 52.2% of the sample was female. 49.7% lived in Flanders, 40.6% in Wallonia and 9.8% in Brussels. 55% of the sample had completed some form of higher education, either university or college. Regarding social class, 49.9% belonged to the top four levels in an eight-level scale. 20.6% of the sample reported that they had personally lived in poverty, either at that moment or at some point in the past.

Participants were randomly allocated to one of the five conditions (i.e, four experimental conditions or benchmark condition). There were no significant differences in gender, language, highest level of education, social class or having personally experienced poverty between the conditions, all χ², p > .140. Also, there was no significant difference between the conditions regarding the age of participants, F(4, 995) = .760, p = .551.

Research design

To ensure the frames’ ecological validity, the experimental conditions were based on the results of an inductive framing analysis on how the Belgian press (n = 103) presented child poverty. An adapted version of that study’s frame matrix is included in Supplementary Online Appendix C. Van Gorp and Gourdin (2015) conducted 16 focus groups with a diverse range of stakeholders, including parents living in poverty, to further ensure that their frames were ecologically valid. Based on the focus groups, two deproblematizing frames and one problematizing frame were selected for application in the stimuli:

• The deproblematizing frame The seed argues that all children – including those living in poverty – have talents and possibilities. Poverty is a problem insofar it hinders children’s development and growth. By investing in poor children now, society will be able to reap the benefits later.

• The deproblematizing frame Solidarity presents the idea that child poverty reminds society of the core values on which it is built. It is sufficient to show solidarity with each other, and then poverty will no longer be a problem for a Western country.

• The problematizing frame A stumbling block stands for the perception that poverty is a persistent hindrance, not only for the children but also for society at large. Society needs to intervene to prevent worse.

A between-subject 2 (Problematizing frame: A stumbling block – none) x 2 (Deproblematizing frame: The seed – Solidarity) posttest-only design was used for the experiment. This was expanded with a benchmark control condition, where participants saw no campaign and received no information on child poverty. This design resulted in two conditions that presented a single deproblematizing frame, and two dual-frame conditions that combined a deproblematizing frame with the problematizing frame, A stumbling block. Two social representations were addressed. First, the social representation of the (poor) child, as innocent and vulnerable, which is difficult to reconcile with personal responsibility for poverty. The archetype of the ‘vulnerable child’ is paired with another social representation of poor children, namely the risks that can arise when they become teenagers: school drop-out, loitering youngsters and drug use. This enumeration is the ‘elephant in the room’ that some test participants were deliberately confronted with and that was strategically kept out of sight for others.

Stimuli

The stimuli were four versions of a print ad campaign that aimed to raise awareness of child poverty (see Supplementary Online Appendix B). Participants were invited to watch the advertisement for at least fifteen seconds, before they were able to proceed. The campaign was designed for the purpose of this study, with the aid of a professional communication agency. At first glance, it seemed to be an advertisement from a financial institution, posing the question: “Are you looking for an investment that yields good long-term results?”. In conditions using the deproblematizing frame The seed, the baseline was “The battle against child poverty: a profitable investment.” In conditions with the deproblematizing frame Solidarity, the baseline was “Contribute and make child poverty go away.”

The body copy contained the main manipulation. There were four versions of similar length, each starting with a factual element: “In Belgium, 1 in 5 children live below the poverty line.” Additionally, all versions ended with “Let us give every opportunity to children living in poverty” and a link to a fictitious website, www.stop-childpoverty.org. The manipulated parts of the body copy read as follows:

1. The seed: “That is why it is essential to invest in day care, good education and quality leisure activities for everyone. By investing in children, we give them the chance to evolve and develop their talents.”

2. Solidarity: “This is despite the fact that children are our capital. By intervening now, we can do away with this injustice. That is why it is high time to take action. Solidarity means helping to realize the dreams of all children”

3. Dual frame: A stumbling block and The seed: “Not intervening will increase the chance of problems in the long term: school dropout, loitering youths, drug addiction, … That is why … [see condition 1]”

4. Dual frame: A stumbling block and Solidarity: “This is despite the fact that children are our capital. Not intervening will increase the chance of problems in the long term: school dropout, loitering youths, drug addiction, … That is why … [see condition 2]”

Measures

The dependent variable in the first research question, namely the effectiveness of a message, can encompass several aspects. In relation to poverty, we elaborate on the renowned framing study by Iyengar (1990) on the attribution of responsibility. To go beyond the message’s non-committal nature, the attitude towards welfare and the willingness to donate were examined as the main dependent variables. In addition to several demographic variables (language, gender, age, educational level and social class), funding allocation to child poverty, beliefs about (in)equality, previous contact with people living in poverty, and attitude towards campaigning served as independent variables. Supplementary Online Appendix A shows a general assessment of participants’ impression of the campaign. Supplementary Appendix G shows the descriptive statistics of the dependent variables and control variables.

Before participants were exposed to a stimulus, they were presented with a funding allocation task that implicitly measured how important they found the problem of child poverty to be (Tagler & Cozzarelli, 2013). Specifically, participants were asked to allocate €1,000,000 of the federal budget among six social policy domains: child poverty, mental healthcare, stimulating employment, public transport, the environment and providing accommodation for refugees. They were instructed to keep in mind how important they personally found each of these domains to be when determining the funding for each of these domains. At this point, participants did not yet know the survey would be about child poverty, as the briefing told them that the research was about their “perception of several issues in society.” For use in the analyses, the allocated amount was divided by 50,000, so that a one-unit increase was associated with an increase of €50,000 in donations.

Personal responsibility was measured using a scale adapted from Corrigan et al. (2003). The items were reworded to measure to what extent participants held the parents and poor children responsible for their own poverty. Two items were added to account for alternative attributions for poverty, namely societal causes and bad luck (Bullock et al., 2003; Lepianka et al., 2010; Tagler & Cozzarelli, 2013). These were respectively: “It is society’s fault that there are poor children in Belgium” and “Most often, it is a matter of bad luck that parents and their children end up in poverty.” A Principal Component Analysis (PCA) with Varimax rotation revealed two factors. However, both had very poor internal consistency (Cronbach’s alpha < .10). By omitting societal attributions, the first factor reached good internal consistency (Cronbach’s alpha = .80) and measured personal blame directed towards the parents. The internal consistency of the second factor did not reach acceptable levels, even after omitting items. Therefore, the three other questions were used as single items for the purposes of the regression analyses. These are societal causes, bad luck, and personal responsibility for living in poverty.

Beliefs about inequality were measured with an 11-item scale developed by Kleugel and Smith (as cited in Bullock et al., 2003). Example questions are: “If incomes were more equal, nothing would motivate people to work hard” and “More equality in incomes would allow my family to live better.” These items were rated on a five-point Likert-scale, ranging from “Completely disagree” to “Completely agree.” A PCA using Varimax rotation revealed two distinct components, beliefs about inequality (Cronbach’s alpha = .87) and beliefs about equality (Cronbach’s alpha = .83). Higher scores on the factor indicated more support for inequality or equality, respectively.

Attitude to campaigning was a three-item scale that asked participants to rate items such as “campaigning is important to make child poverty a subject of discussion” on a five-point Likert scale, ranging from “Completely disagree” to “Completely agree” (Van Gorp et al., 2012). After reverse coding scores on “campaigning is useless because it has no effect,” a PCA found a single factor, with good internal consistency (Cronbach’s alpha = .81). The higher a person scored, the more they agreed that campaigning is important.

Seeing as previous contact with people in poverty is likely an important moderator for framing effects, the survey used nine statements to gauge previous contact. The first seven questions were adapted from Corrigan et al. (2003) and included such items as “I have observed persons living in poverty on a frequent basis.” For the purpose of this study, the item, “I work as a volunteer, which brings me in contact with people living in poverty,” was added. Each statement was rated as either 0 (no) or 1 (yes). These eight items were added up. Higher scores indicated more previous contact with people living in poverty.

Finally, to measure personal experience with poverty, participants were also asked whether they were living or had lived in poverty. This variable was added as a separate binary predictor in the analysis.

Welfare attitudes was measured using a four-item scale by Gilens (1996) (e.g., “Most people on welfare could get by without it if they really tried”). Participants rated their agreement with the statements on a five-point Likert-scale, from “Completely disagree” to “Completely agree.” After reverse coding the two items that measured negative attitudes towards welfare, a PCA identified one factor, which yielded sufficient internal consistency (Cronbach’s alpha = .67). Higher scores on the factor indicated more support for welfare.

Since positive attitudes towards poor persons do not adequately predict behavior (Tagler & Cozzarelli, 2013), the final question gauged donating behaviors. Participants were informed that they could win one of five €50 vouchers. They had to indicate whether they wanted to receive a voucher, or if they wanted their prize to be donated to charity. In other words, participants had to choose whether they wanted to keep the money for themselves, or if they wanted it to be donated to help others. Three of the five winners wanted to donate their prize.

Results

Attribution of responsibility

The stimuli yielded significant explanatory power in the linear regression analysis that studied the relationship between the applied frames and societal attributions for poverty, F (15, 984) = 23.03, p < .001 (see Table 1). All independent variables were entered in the same block, and there were no problems with multicollinearity. The model was able to explain 24.9% of the variability. The deproblematizing frame Solidarity, when presented on its own, led to a significant increase in agreement that society is to blame for child poverty (β = 0.11, p = .005), supporting H1a. The predictive power of the deproblematizing frame The seed approached significance (β = 0.07, p = .066). The most powerful predictor was beliefs about equality: those who felt that society ought to become more equal, placed more blame on society for the existence of child poverty (β = 0.27, p < .001).

Table 1. Regression analysis on societal responsibility for poverty, bad luck as a cause, and attitudes towards welfare (N = 1000)

	Societal responsibility for poverty			Bad luck as a cause for poverty			The attitudes towards welfare
Predictor	B	β	p	B	β	p	B	β	p
Language: French	0.66	0.15	***	−1.25	−0.33	***	−0.18	−0.09	***
Gender: Female	−0.20	−0.05		0.18	0.05		−0.07	−0.03
Age	−0.01	−0.08	**	0.00	−0.02		0.00	0.07	*
Degree	−0.01	0.00		0.05	0.06		0.07	0.14	***
Social class	0.05	0.05		−0.01	−0.02		0.03	0.08	*
Funding allocated to child poverty	0.07	0.08	**	0.02	0.03		0.00	−0.01
Previous contact with poverty	0.11	0.09	**	0.04	0.04		0.06	0.10	***
Personal experience of poverty	−0.23	−0.04		−0.33	−0.07	*	0.16	0.07	*
Support for inequality	−0.46	−0.22	***	0.00	0.00		−0.32	−0.32	***
Support for equality	0.59	0.27	***	0.17	0.09	**	0.22	0.22	***
Attitude to campaigning	0.30	0.14	***	0.00	0.00		0.24	0.24	***
Dual frame: The seed + A stumbling block	0.20	0.04		−0.17	−0.04		0.13	0.05
Dual frame: Solidarity + A stumbling block	0.07	0.01		−0.52	−0.11	**	0.05	0.02
Single deproblematizing frame: The seed	0.36	0.07		−0.31	−0.07		0.10	0.04
Single deproblematizing frame: Solidarity	0.56	0.11	**	−0.20	−0.04		0.29	0.12	**
Constant	4.71		***	4.17		***	−0.84		***
N	1000			1000			1000
Adj. R^2	0.25			.12			0.31

* p < .05; ** p < .01; *** p < .001

The frames were also significant predictors in the regression analysis on the attribution of poverty to bad luck, F (15, 984) = 10.31, p < .001. Once again, all independent variables were entered at the same time, and there were no issues with multicollinearity. The analysis explained 12.3% of variability. In contrast to the prediction made in H1a, the combination of the problematizing frame A stumbling block and Solidarity led to a significant decrease in agreement that it is just a matter of bad luck that leads people into poverty (β = −0.11, p = .005). Language was the most powerful predictor, with the French-speaking half of the sample agreeing less with the statement that it is just a matter of bad luck that people become poor (β = −0.33, p < .001).

A next linear regression analysis was performed to gauge the predictive power of the frames for attitudes towards welfare, F (15, 984) = 31.41, p < .001. The independent variables were all entered in the same block, and there were no problems with multicollinearity. The analysis was able to explain 31.3% of the variability. Supporting H1b, the deproblematizing frame Solidarity was a significant predictor, leading to more positive attitudes towards welfare, β = 0.12, p = .001. The strongest predictor, however, was the beliefs about inequality. Respondents who had favorable attitudes towards inequality were significantly less positive about welfare, β = −0.32, p < .001.

When introducing interaction effects with personal experience with poverty (see tables in Supplementary Online Appendix D–F, and visualizations in Supplementary Online Appendix H), the analyses clearly indicated that only people who have never personally experienced poverty were significantly affected. Specifically, for societal responsibility there were two significant simple main effects: The seed, B = 0.47, β = 0.09, p = .034, and Solidarity, B = 0.44, β = 0.08, p = .047. For bad luck as a cause of poverty, the combination of Solidarity with A stumbling block had a significant simple main effect, B = −0.51, β = −0.11, p = .015. For attitudes towards welfare, Solidarity had a significant main effect B = 0.28, β = 0.11, p = .005. None of the interaction terms were significant, indicating that the stimuli had no significant effects on people who have themselves experienced poverty. This is in line with the predictions made in H2a and H2b.

Donating behavior

The final question in the survey informed participants that they had a chance to win €50, which they could either keep for themselves or donate to charity. In total, 53.7% of participants chose to donate their prize.

The data was split depending on whether the participants had personally lived in poverty (see Supplementary Online Appendix H for a visualization). When considering the people without personal experience of poverty, the analysis showed that 50.0% of the control group chose to donate, whereas 64.9% of those who saw the campaign featuring just Solidarity chose to back a charity with their money. This result (+14.9%) is a clear effect of the deproblematizing frame Solidarity, albeit only for people who did not have any personal experience with poverty. For people with personal experience of poverty, 72.5% of people who been exposed to the stimulus containing The seed chose to receive a voucher.

A logistic regression analysis reiterates that deproblematizing frames affected people with and without a personal history of living in poverty differently (see Table 2). Supporting H1c, people without personal experience were 1.64 times more likely to donate their money to charity after being exposed to Solidarity, p = .047. H2c predicted that framing would have less of an effect on the willingness to donate by people with personal experience. However, this prediction was not borne out, as two deproblematizing frames had significant effects: People with a personal experience of poverty were less likely to donate to charity after exposure to a campaign containing either The seed (OR = 0.28, p = .021) or Solidarity (OR = 0.31, p = .031).

Table 2. Logistic regression analysis on donating behavior (N = 999)

Predictor	B	OR	p	B	OR	p
Constant	0.20	1.22		0.02	1.02
Language: French	−0.56	0.57	***	−0.53	0.59	***
Age	0.01	1.01	*	0.01	1.01	*
Social class	−0.11	0.90	***	−0.10	0.90	***
Funding allocated to child poverty	0.07	1.07	*	0.07	1.07	*
Personal experience of poverty	−0.16	0.85		0.43	1.54
Attitude to campaigning	0.34	1.41	***	0.35	1.41	***
Complementary frame: The seed + A stumbling block	0.04	1.04		0.14	1.15
Complementary frame: Solidarity + A stumbling block	−0.10	0.90		−0.05	0.95
Single deproblematizing frame: The seed	−0.21	0.81		0.05	1.05
Single deproblematizing frame: Solidarity	0.25	1.28		0.50	1.64	*
Interaction effects
Experience x The seed + A stumbling block				−0.43	0.65
Experience x Solidarity + A stumbling block				−0.19	0.83
Experience x The seed				−1.27	0.28	*
Experience x Solidarity				−1.18	0.31	*
N	999			999
Cox & Snell R^2	0.08			0.09
Nagelkerke R^2	0.11			0.12

Discussion

This study’s main aim was to contribute to framing theory, a concept that otherwise risks disappearing into a metaphorical substitute for representation (Scheufele & Iyengar, 2017). In answer to the first research question, which related to the potential contribution of deproblematization to a message’s effectiveness, it can be concluded that the applied combination of frames led to an increase in a societal responsibility for poverty and a decrease in individual responsibility. The assumption that deproblematizing frames can produce powerful effects was most pronounced from the finding that the participants were significantly more willing to donate if they were confronted with a deproblematizing frame than if they did not see any campaign (+14.9%). Furthermore – and perhaps related – they had more positive attitudes towards social welfare. These results indicate that a campaign using the deproblematizing frame Solidarity would be effective in changing both attitudes and behavior regarding the battle against child poverty. Furthermore, a reference to a problematizing frame appears to outweigh the beneficial effect of the deproblematizing frame. Finally, the study confirmed the insight described in literature that framing is less effective among people with high involvement (e.g., Donovan & Jalleh, 2000; Luong et al., 2019).

This study into FCT builds on previous research in four key ways. First, there is the typology of issue frames, according to whether the frame contributes to a problematizing definition of an issue. Without lapsing into postmodern relativism, the research showed that there are different perspectives in looking at an issue that goes beyond a black (negative) and white (positive) representation of social reality. In the case of poverty, a problematizing frame points to the responsibility of the person living in poverty, which may be stigmatizing. A deproblematizing frame that acts as a counter-frame, on the other hand, decreases the problematizing nature of the definition and shifts attention to a collective responsibility. In this way, FCT is also an addition to valence framing research (e.g., Han & Wang, 2015). Second, because deproblematizing frames represent unexpected approaches to a global society, they can enrich a framing analysis of societal issues (see also Aklin & Urpelainen, 2013). Therefore, FCT provides an insight into as well as a way out of paradoxical definitions of societal issues, for instance, when a definition risks strengthening the stigma or obscures the sheer seriousness of the issue. As such, FCT can emphasize the societal relevance of research into communication and framing. Third, framing effects of deproblematizing frames can be particularly strong, but the insertion of a problematizing frame can cause that effect to disappear. This underlines that framing is not a static process, but has dynamic qualities (see Reese, 2007, p. 150). In child poverty, the child itself represents a compelling deproblematizing frame, with clear positive associations. Even though teenagers are children, people seemingly do not spontaneously think about school dropout and drug use in the context of child poverty. Although the campaign explicitly stated that these downsides would be avoided, the accompanying schemata overshot the intended effects. Fourth, framing is about rather minor differences in a text that can cause a major impact. Adding a single sentence that evokes a specific mental image can both cause effects and make them disappear. This is a very pertinent finding from a strategic point of view, because referencing negative aspects of an issue that one explicitly wants to avoid can cause the public to take the message further into consideration and come to an opposite conclusion than originally intended. Organizations may therefore fail to achieve their objectives.

The findings of this research are likewise relevant to the practice of campaign development, the strengthening of social awareness of poverty, and the strategic communication in social issues management (Coombs & Holladay, 2018). The approach proposed here emphasizes the importance for communicators to consider different perspectives of looking at social reality when formulating messages. For organizations, this requires a high level of awareness regarding their own framing as well as insight into the frames used by the various stakeholders and finding a way to respond to the frames of specific audiences.

In conclusion, FCT postulates that: (1) in common public discourse, frames that define issues as problematic dominate; (2) frames that define issues as non-problematic or at least as less problematic can function as counter-frames; (3) deproblematizing frames can produce powerful effects when applied in isolation in a message and (4) personal involvement in the issue is low; (5) however, a (minimal) reference to a problematizing frame can cause this effect to disappear, (6) and does so especially when it represents an idea that people do not spontaneously think of. This conceptualization can also be seen as an extension of both gain-loss framing, because it is not limited solely to risk estimates; and of prognostic and diagnostic framing from SMT (e.g., Gahan & Pekarek, 2013). After all, according to the conceptualization used here, every frame involves a prognostic and diagnostic component.

The study also has limitations. For example, the precise interplay between the reasoning devices, i.e., cause, problem definition, and solution, deserves more attention. The same goes for the emotionality associated with it. For that, more subjects and more stimuli need to be examined. There also is more to explore in the context of FTC. Firstly, there should be ways to gain efficient insight into counter-frames that are able to act as a counterbalance to the usual dominant problematizing framing. Secondly, the stimuli should be better attuned to the segmentation and divergent profiles of the participants. What works for one participant can lead to an opposite effect for another. The structures of the composed stimuli were not equivalent (see Scheufele & Iyengar, 2017), and therefore the specific choice of wording may determine the results. Thirdly, the methodology for the psychological processing of stimuli should be refined. Survey experiments also have clear shortcomings. For example, if some words can already cause a significant impact, this presents the challenge of setting up experiments in which even the formulation of survey questions can already have an influence on the answers. Developing more sophisticated methodologies to test framing effects is certainly a challenge. Furthermore, the nature of frames, whether they function as a problematizing or a deproblematizing frame deserves more attention, as not all frames have been created equally.

Ethics

Ethical approval for this study was obtained from the Social and Ethics Committee (SMEC), approval G-201512415. Informed consent was obtained from all individual participants involved in the study.

Disclosure statement

The authors do not report any conflict of interest. The data that support the findings of this study are available on request from the corresponding author, BVG. The data are not publicly available because they contain information that could compromise the privacy of the research participants.

Supplemental material

Supplemental data for this article can be accessed on the publisher’s website.

Additional information

Funding

This work was supported by the King Baudouin Foundation and Research Foundation Flanders (FWO), grant number G070416N.